CiteFetch VS CiteWise.dev

CiteWise.dev's answer:

CiteWise checks whether a reference points to a real scholarly work, and it shows the evidence behind each result. It doesn’t format citations. For every reference you see which fields matched (title, authors, year, publication, DOI), which conflicted, and which source records each result came from. Matching is done by deterministic, field-by-field scoring against Crossref, DataCite, OpenAlex, Semantic Scholar, PubMed and arXiv. An LLM only helps parse messy input and never decides whether a reference is real.

CiteWise.dev's answer:

Most citation tools either generate formatted references or give a yes/no “fake or real” answer. CiteWise gives you the evidence and leaves the decision to you: field-level matches, metadata and DOI conflicts, check completeness, and publication updates such as retractions where available. You can audit a whole bibliography (paste it, or import BibTeX, RIS, EndNote or CSL-JSON), record review decisions, and export HTML, CSV or JSON reports. A free account includes 10 checks per month.

CiteWise.dev's answer:

Researchers, graduate students, librarians and academic editors who need to confirm a bibliography is accurate before submission. This includes people checking references suggested by ChatGPT or other AI writing tools, and anyone who has to verify a long reference list without opening every DOI by hand.

CiteWise.dev's answer:

AI writing tools made it easy to produce references that look correct but point to the wrong paper, or to no paper at all. Checking them by hand means resolving DOIs and comparing titles, authors and years one at a time. I built CiteWise to automate that lookup while keeping the evidence visible. Early on I tried letting an LLM judge whether a reference was real and dropped the idea: LLMs are good at producing things that look right, which is the same problem that creates fake citations. CiteWise now uses deterministic matching against scholarly databases, and the LLM is limited to helping parse messy input.

CiteWise.dev's answer:

Java 17 and Spring Boot 3 on the backend, Next.js 15 on the frontend, PostgreSQL on Supabase, Clerk for authentication and Creem for payments. Reference data comes from the Crossref, DataCite, OpenAlex, Semantic Scholar, PubMed and arXiv APIs.

CiteWise.dev's answer:

CiteWise launched in November 2026 and doesn’t have named customers yet Early users are individual researchers and students

A few notes. I didn’t make up any customer names for the last question, because you have about 8 registered users and none of them are paying yet. Listing made-up institutions on a directory site is easy to get caught on, and it damages credibility. If the form won’t accept that answer, you can leave it blank. The story answer is a draft based on the LLM bug you wrote about in your dev.to post, so replace it with your real reasons for starting the project, like a personal experience or what your classmates went through. Specific details make it more convincing.